Exploring forecasting models for tourist arrivals in international tourist hotels

Accurate tourism forecasting is particularly crucial not only for governments and practitioners but also for investor resource allocation and decision making. The main objective of this study was to obtain more accurate forecasts of tourist arrivals for a specific international tourist hotel in Taiwan by comparing the autoregressive integrated moving average (ARIMA) model and the unrestricted vector autoregressive (VAR) model, which have rarely been employed in the hotel industry. Monthly data covering the period 2000 M1 to 2010 M12 were collected from the Monthly Report on Tourism, which is published by the Tourism Bureau of Taiwan. The studied variables included the gross domestic product (GDP), the consumer price index (CPI), the exchange rate, and the hotel operating characteristics. Forecasting performance is assessed in terms of mean absolute per centage error (MAPE). The superior performance of the VAR model implies that the inclusion of endogenous variables is required in forecasting international tourist hotel demand. An impulse response analysis was performed to assess the impact level of tourist arrivals in response to shocks in the economic variables in the VAR model.